Experimental Research Of Damage Detection On Concrete Arch Dam Based On The Vibration

In China, thousands of dams have been constructed in past years. Damage may occur in the period of service of each concrete dam. It will influence the design function of hydropower station and seriously it will lead to the dam to destroy and bring disaster to the lives and properties of the people in the lower reaches, if these damages develop arbitrarily and can't be found timely. Therefore to these hydro-structure, the damage analysis and safety assessment is necessary.In this paper,firstly,the writer imaginarily designs a real single-curvature arch dam. According to similarity theory,model is then made and test in the experimental laboratory.Undamage model is not only simulated,but also crack damage model,which is loaded from jacking apparatus,is mocked,and dynamic characteristics, including primiarily frequency and modal, is tested in all kind of statuses completely.Analysis indicates apparents that crack damage take effect on stiffness change and serious damage lead to more change. Secondly, numerical model of the real single-curvature arch dam is made thought ANSYS-finite element software.Undamage numerical model and crack damage numerical models ,compared with experimental models ,is correct and proper.Three damage numerical models are analysed with modal and give the ten natural frequencies and modes.From the results,numerical model and experimental model come to the same conclusion.At last, dam damage can be detected by a new approach called statistical neural network , based on the neural network and statistical principle. The conception of possibility is applied in the sequence to represent the occurrence of structural damage. Because statistical neural network not other neural networks has noise-immune ability ,experimental model and numerical model with the added-noise results can be correctly detected.The results show that it is possible to practically detect the damage in concrete arch dams by testing dynamic characteristics change and using statistical neural network . Of course, the method is limited by environmental factor and precision of instrument when used in practice.